Comparing GOSAT and OCO-2 Observations with High-resolution CO2 Simulation Driven by Fluxes Optimized with Surface Data Inversion

Abstract:

We apply a high-resolution inverse modeling system NTFVAR based on coupled NIES-TM-FLEXPART transport model to optimize surface CO₂ fluxes during 2009-2018 using Obspack dataset. The optimized fluxes are used to simulate column average CO₂ (X_CO2) observed by GOSAT (NIES Level 2 dataset) during 2009-2019. The simulated vertical profiles are confirmed to match the independent aircraft observations in free troposphere within 0.5 ppm. The differences between optimized simulation and GOSAT are analyzed as seasonal averages for each 10-degree latitudinal band and are used to estimate the mean deviations depending on season and latitude separately for each mode of observation: nadir, glint, and gain level. The estimated differences can be used to extend spatially the estimates of GOSAT X_CO2 retrieval biases observed at TCCON network. Model to observation comparison for OCO-2 (V9) data for 2014-2019 is also prepared. OCO-2 data is combined to 1-second averages matching the transport model resolution of 0.1 degrees. Monthly mean differences by latitudinal bands are produced to serve as input to bias correction for use in inverse modeling.